The study of generative grammar has undergone substantial theoretical refinement since its inception in the 1950s. One of the most notable shifts concerns the movement away from the concept of Deep Structure, prominent in Chomsky’s earlier frameworks, toward the operation of Merge, which occupies a central position in the Minimalist Program. This essay examines that transition, outlining the original role of Deep Structures, the reasons for their eventual abandonment, and the emergence of Merge as a simpler, more fundamental mechanism. It draws on key texts to demonstrate how these developments reflect broader efforts to achieve greater explanatory adequacy while reducing theoretical complexity.
Deep Structure in the Standard and Extended Standard Theories
In the model presented in Aspects of the Theory of Syntax, Chomsky (1965) proposed that sentences are generated through the interaction of a base component, consisting of phrase-structure rules, and a transformational component. Deep Structure was defined as the level at which the underlying grammatical relations are represented prior to the application of transformations. This level was considered crucial for semantic interpretation because it encoded the information necessary to determine meaning before surface adjustments occurred.
The separation between Deep Structure and Surface Structure allowed for systematic accounts of phenomena such as passivisation and question formation. For instance, the active sentence “The boy kicked the ball” and its passive counterpart “The ball was kicked by the boy” were said to share the same Deep Structure, with transformations producing the surface differences. This architecture provided a clear division between the generative base and subsequent operations, yet it also introduced considerable machinery in the form of both phrase-structure rules and a range of transformations.
Challenges and the Move toward Minimalism
Subsequent work revealed several difficulties with the Deep Structure model. The proliferation of transformations risked undermining the goal of explanatory adequacy, as learners would need to acquire a large and language-specific set of rules. In addition, questions arose concerning the precise location of semantic interpretation. Jackendoff (1972) and others argued that surface structure also contributed to meaning, weakening the claim that Deep Structure alone determined semantic relations.
Chomsky’s own later writings indicate a growing preference for theoretical parsimony. By the late 1980s and early 1990s, the Principles and Parameters framework had already reduced much of the earlier descriptive apparatus. However, residual elements, including the distinction between Deep Structure and Surface Structure, continued to be viewed as potentially unnecessary. The Minimalist Program, first outlined in Chomsky (1995), sought to eliminate levels of representation that lacked strong empirical or conceptual motivation.
The Introduction of Merge
Within the Minimalist framework, the operation Merge replaces the earlier phrase-structure component and many transformations. Merge is a simple, recursive procedure that takes two syntactic objects and forms a new object from them. It operates under the requirement that each application must be driven by feature checking, ensuring that structure building is economical.
Unlike Deep Structure, which presupposed a pre-existing set of phrase-structure rules, Merge generates hierarchical structure incrementally. Binary Merge first combines two lexical items; subsequent applications can combine a lexical item with an existing syntactic object or two complex objects. This recursive property accounts for the unbounded nature of natural language without requiring separate levels of representation. Furthermore, the distinction between Deep Structure and Surface Structure is removed; instead, structures are transferred to the interfaces (PF and LF) at appropriate points in the derivation.
The shift also affects how grammatical relations are captured. In the earlier model, relations were encoded at Deep Structure; in minimalist terms, they arise through the application of Merge and subsequent movement operations driven by feature valuation. This change streamlines the grammar by reducing the number of stipulated levels while preserving the capacity to explain displacement phenomena through internal Merge, formerly understood as movement.
Comparative Implications
The replacement of Deep Structure by Merge carries several theoretical consequences. First, it aligns syntax more closely with requirements of learnability, since Merge is a single, universal operation rather than a collection of language-particular rules. Second, it emphasises interface conditions: the grammar is evaluated according to how efficiently it satisfies the needs of the sound and meaning systems.
Nevertheless, some scholars have questioned whether the minimalist reduction truly eliminates all descriptive devices or merely relocates them. For example, the need for feature specifications and selectional restrictions might be seen as preserving a degree of complexity comparable to earlier phrase-structure rules. Even so, the conceptual motivation for Merge—that it constitutes the minimal apparatus sufficient for structure building—has influenced much subsequent research in both theoretical syntax and language acquisition studies.
In summary, the transition from Deep Structure to Merge marks a deliberate move toward theoretical minimalism. While Deep Structure provided an explicit level for representing underlying relations, Merge achieves similar descriptive coverage through a single recursive operation applied at the interfaces. This development illustrates the field’s continued effort to balance empirical coverage with conceptual simplicity.
References
- Chomsky, N. (1965) Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
- Chomsky, N. (1995) The Minimalist Program. Cambridge, MA: MIT Press.
- Jackendoff, R. (1972) Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press.
